The Research Of International Crude Oil Price Forecasting Model

This paper adopts between December 2000 and July 2015 WTI crude oil futures prices monthly data as the research object.Qualitative analysis part, first of all a retrospect of the common international commodity futures pricing theory and crude oil price forecasting research present situation, and then summarized from the aspects of economic financial fundamentals and the main factors influencing the long-term and short-term price fluctuations of crude oil, to the back of the empirical analysis to determine the model input variables provide logical basis.Empirical analysis, first of all, ARIMA,traditional multiple linear regression, artificial neural network is widely used, such as the principle, advantages and limitations of single model, and then combining the multiple linear regression model and ARMA model constructing portfolio model, through the sample prediction, and the traditional multiple linear regression model and ARIMA model to predict the effect of comparative analysis. The empirical results show that the combination model of the root mean square prediction error is smaller than that of ARIMA model and the traditional multiple linear regression model, compared with better prediction effect.